针对现有的柑橘检测算法准确率低、模型参数量大、检测实时性差、不适用移动采摘设备等问题,提出一种基于改进轻量模型YOLO-DoC的柑橘检测方法。引入Bottleneck结构的ShuffleNetV2网络作为YOLOv5骨干网络模型,构造轻量化网络。同时加入...针对现有的柑橘检测算法准确率低、模型参数量大、检测实时性差、不适用移动采摘设备等问题,提出一种基于改进轻量模型YOLO-DoC的柑橘检测方法。引入Bottleneck结构的ShuffleNetV2网络作为YOLOv5骨干网络模型,构造轻量化网络。同时加入无参型SimAM注意力机制提高复杂环境下对目标的识别精度。为了提高检测网络对于目标果实的边界框定位精度,通过引入Alpha-IoU边界框回归损失函数的方法来获取目标的边界框。实验显示,YOLO-DoC模型的P(precision)值和mAP(mean average precision)值分别为98.8%和99.1%,参数量缩减为YOLOv5网络的1/7,模型的大小为2.8 MB。改进后的模型相比于原网络模型具有识别速度快、定位准度高以及占用内存少的优势,在满足精准采摘工作要求的前提下可以提高采摘效率。展开更多
China is one of the countries which were affected very seriously by tropical cyclones. In this paper, the active features of tropical cyclones which affected and landfell over China, the distinguishing characteristics...China is one of the countries which were affected very seriously by tropical cyclones. In this paper, the active features of tropical cyclones which affected and landfell over China, the distinguishing characteristics of typhoon disasters and related formation laws have been studied, and the impact of trphoon disasters on China's socio-economic development have been discussed. The study reveals that typhoon disasters which affected China are characterized by high freqency in occurrence, violent in sudden occurrence, remarkable effects in China,widescope in impact and severe intensities in disasters, and this type of disaster was caused by huge gales, rainstorms, storm surges and reatal chain-disasters.Typhoon disasters have not only caused great casualties but also exerted severe effects on every economic departments in China. There is an annual increasing tendency in direct economic loss caused by taphoon disasters and since 1990, the annual average economic loss has totalled up to over lo billion RMB yuan.展开更多
为评估罗非鱼湖病毒(tilapia lake virus,TiLV)跨境传入风险,按照OIE传入风险评估框架,从危害识别、风险评估及风险管理3个方面,开展了TiLV跨境传入风险分析,并提出了相应的风险控制措施。分析认为:活的敏感鱼类(包括亲鱼、鱼卵和鱼苗)...为评估罗非鱼湖病毒(tilapia lake virus,TiLV)跨境传入风险,按照OIE传入风险评估框架,从危害识别、风险评估及风险管理3个方面,开展了TiLV跨境传入风险分析,并提出了相应的风险控制措施。分析认为:活的敏感鱼类(包括亲鱼、鱼卵和鱼苗),用做鲜活饵料的鱼类(野杂鱼)及冰冻和冰鲜的敏感鱼类(包括去除内脏的鱼肉)TiLV传入风险较高,而敏感鱼加工类产品、非敏感鱼类(包括活的和其他形式的产品)及运输活鱼的水、包装、运输工具和用具等传入风险较低。建议不从疫区进口高风险产品,对于低风险产品可以自由贸易,而对运输活鱼的水、包装、运输工具和用具等要进行强制常规消毒。展开更多
文摘针对现有的柑橘检测算法准确率低、模型参数量大、检测实时性差、不适用移动采摘设备等问题,提出一种基于改进轻量模型YOLO-DoC的柑橘检测方法。引入Bottleneck结构的ShuffleNetV2网络作为YOLOv5骨干网络模型,构造轻量化网络。同时加入无参型SimAM注意力机制提高复杂环境下对目标的识别精度。为了提高检测网络对于目标果实的边界框定位精度,通过引入Alpha-IoU边界框回归损失函数的方法来获取目标的边界框。实验显示,YOLO-DoC模型的P(precision)值和mAP(mean average precision)值分别为98.8%和99.1%,参数量缩减为YOLOv5网络的1/7,模型的大小为2.8 MB。改进后的模型相比于原网络模型具有识别速度快、定位准度高以及占用内存少的优势,在满足精准采摘工作要求的前提下可以提高采摘效率。
文摘China is one of the countries which were affected very seriously by tropical cyclones. In this paper, the active features of tropical cyclones which affected and landfell over China, the distinguishing characteristics of typhoon disasters and related formation laws have been studied, and the impact of trphoon disasters on China's socio-economic development have been discussed. The study reveals that typhoon disasters which affected China are characterized by high freqency in occurrence, violent in sudden occurrence, remarkable effects in China,widescope in impact and severe intensities in disasters, and this type of disaster was caused by huge gales, rainstorms, storm surges and reatal chain-disasters.Typhoon disasters have not only caused great casualties but also exerted severe effects on every economic departments in China. There is an annual increasing tendency in direct economic loss caused by taphoon disasters and since 1990, the annual average economic loss has totalled up to over lo billion RMB yuan.
文摘为评估罗非鱼湖病毒(tilapia lake virus,TiLV)跨境传入风险,按照OIE传入风险评估框架,从危害识别、风险评估及风险管理3个方面,开展了TiLV跨境传入风险分析,并提出了相应的风险控制措施。分析认为:活的敏感鱼类(包括亲鱼、鱼卵和鱼苗),用做鲜活饵料的鱼类(野杂鱼)及冰冻和冰鲜的敏感鱼类(包括去除内脏的鱼肉)TiLV传入风险较高,而敏感鱼加工类产品、非敏感鱼类(包括活的和其他形式的产品)及运输活鱼的水、包装、运输工具和用具等传入风险较低。建议不从疫区进口高风险产品,对于低风险产品可以自由贸易,而对运输活鱼的水、包装、运输工具和用具等要进行强制常规消毒。
